Exploring How Regulatory Sandbox Act as an Institutional Catalyst for AI-Driven Business Model Innovation

Abstract

Regulatory sandboxes have emerged as adaptive policy instruments that enable firms to test innovative technologies under controlled conditions while allowing regulators to observe, learn, and refine governance frameworks. This study conceptualizes the regulatory sandbox as an institutional catalyst for AI–driven business model innovation, emphasizing its dual role in promoting experimentation and institutional learning. Drawing on institutional theory and dynamic capabilities theory, the research develops a comparative framework that explains how sandbox design and governance structures shape organizational learning, legitimacy and capability building. Using secondary data from policy reports, institutional documents,and international databases, the study examines three national cases: the UK, Japan, and Kenya, each representing distinct levels of institutional maturity, governance orientation and development priorities. The cross case analysis shows that while all sandboxes aim to balance innovation and regulation. The UK’s collaborative model emphasizes ethical governance, Japan’s centralized framework aligns sandboxing with industrial strategy and Kenya’s inclusive approach integrates capacity building and digital inclusion. These findings extend theoretical understanding by showing that regulatory sandboxes function not only as legal tools but also as dynamic institutional mechanisms that embed learning, flexibility and legitimacy within governance systems. The study provides practical guidance for policymakers seeking to design sandbox frameworks that promote responsible AI experimentation and support institutional adaptation across diverse economic contexts.